Organizations face challenges in how to manage new infrastructure and data that come with IoT. Ensuring data integrity, security, connectivity and data uploads and updates are concerns, as are issues with vendor management in a field where vendors can be less IT-oriented. What best practices are emerging for IoT and how do these practices impact traditional data, asset and network management?
Here are eight best practices:
Develop a formal edge computing strategy. In 2018, a majority of companies opted to take a slow and measured approach to IoT adoption. This was prudent because it allowed organizations to trial IoT technology and determine “best fit” scenarios for the concept. Unfortunately, the disadvantage of isolated trialing and implementation was that IoT projects were often piecemeal. Given this, 2019 is the year for IT to revisit its IT architecture for the entire enterprise, with an eye toward incorporating IoT and other types of edge computing technology and approaches into the total corporate IT infrastructure. Part of this work should include establishing asset management software and a methodology that can track and trace every IoT device and system that gets deployed at the edge.
Partner with your end users. For businesses with many distributed facilities in different regions of the country or world, it is physically challenging for IT to monitor every piece of IoT computing at the edge. Also, it’s your users (e.g., a manufacturing supervisor running robotics on a production line) who know the business performance needs of the IoT best. This is what makes it imperative for IT to develop open and highly collaborative relationships with the end users who will be the ones running localized edge computing applications like IoT. An open and collaborative partnership between IT and end users can help reduce the adverse impacts of shadow IT, which gets installed by users without IT knowing about it. A strong IT-end user partnership also reduces the probability of a security breach, since IT is part of a collaborative team and can help assure that corporate security is practiced on the edge.
Build zero trust networks. Security and governance can be the most difficult concepts to impress upon end users at the edge, because they are focused on the business and not IT. To address security and governance as seamlessly as possible, organizations are moving to zero trust networks that automatically verify IP addresses and authenticate users from both inside and outside corporate walls. Since no one gains admission to the network until all security criteria have been met, IT can set the network to enforce the necessary user permissions that it defines with end user managers without having to pay a visit or pick up a phone when a user-related network anomaly is detected.
Vet incoming solutions for interoperability. While there has been talk about open systems and standardization for IoT, much of today’s IoT consists of disparate IoT devices created by a broad range of manufacturers. In many cases, these IoT devices and solutions have proprietary operating systems that neither IT nor other vendors know anything about. This complicates connectivity and integration. While technology integration and data exchanges might not be immediate problems as IoT gets introduced on a small scale, the more that IoT gets implemented in the enterprise, the greater the call will be for interoperability of that individual IoT network with other IoT and corporate legacy systems. IT can help itself and the company if it inserts itself early into the IoT decision making process — before an IoT solution is purchased. In this way, a prospective IoT solution can be vetted for scalability and interoperability as well as for features and functions.
Revise your DR and failover plans for IoT. In 2018, 52% of consumers were using IoT devices, and 64% had experienced failures. The picture isn’t much different in business when a manufacturer sees a production line go down because of an IoT glitch in a piece of equipment, or when a sensor fails on a refrigerated truck and a load of lettuce spoils. The situation worsens when an IoT failure shuts down an entire energy grid. All of this boils down to one thing: IT must incorporate new IoT technology into its existing DR plans and test for IoT system health on a regular basis.
Develop realtime IoT data elimination strategies. To date, IT has been reluctant to establish data retention (and elimination) strategies for the legions of big data streaming into companies through IoT and edge computing. One way to edit your data while it is streaming in is to use tools that enable you to set the business rules for the data you want to accept into your system, and what you want to leave out. These tools then analyze the incoming edge/IoT data and eliminate unwanted data before it ever enters your network. The process saves time, and eliminates the need for many follow-up ETL operations.
Revise business processes for man-machine workloads. What IoT at the edge brings to companies is process automation, plus present and predictive analytics and decision making. In many cases, this is offsetting some of the functions humans used to do, but it doesn’t eliminate the need for humans. As this unfolds, a cultural as well as a process shift is occurring. This makes it critical for IT, HR and the business to define the dynamics of human-machine workflows and then retrain personnel so they understand and feel confident in revised roles.
“The US Air Force partnered with Lockheed Martin Skunk Works to demonstrate manned/unmanned teaming to enable rapid action during combat,” said Mark Cole, business strategy and development ISR and UAS, Lockheed Martin Skunk Works. “The experiment, called Have Raider, was designed to demonstrate the technologies required for an unmanned vehicle to fly as a teammate with a manned vehicle in the battlespace.”
Plan for bandwidth constraints. Edge computing almost always involves managing bandwidth constraints. These bandwidth constraints exist on public Internet, and can become an expensive proposition if you decide to implement your own private communication network.
A best practice for many organizations is to establish local processing and storage at the edges of the enterprise where edge computing and the IoT are being used. IT then periodically uploads this data to temporary cloud platforms and ultimately, to central data center storage. Once data is collected in central storage, analytics on the entire body of data can be run to produce the most inclusive and holistic picture of the business. The key to the strategy is developing an efficient batch processing workflow that takes advantage of light traffic bandwidth times to effect the data transports.